chatgpt Timeline - FediScanner

26.08.2025 19:44
tugatech (@tugatech@masto.pt)

OpenAI enfrenta primeiro processo por homicídio culposo devido ao papel do ChatGPT em suicídio
🔗 https://tugatech.com.pt/t71042-openai-enfrenta-primeiro-processo-por-homicidio-culposo-devido-ao-papel-do-chatgpt-em-suicidio

#ChatGPT #judicial #OpenAI #segurança

Show Original Post

Report

26.08.2025 19:40
us (@us@newsbeep.org)

Nightmare rise of AI psychosis. It’s leading to divorce, institutionalisation and even suicide. For Allan it started by asking chatbot a simple question… and nearly broke him. Now he’s overcome his shame to speak out…

As with so many of ChatGPT’s 800million users, Allan Brooks’s experience began when he asked a simple qu…
#NewsBeep #News #US #USA #UnitedStates #UnitedStatesOfAmerica #Mentalhealth #ChatGPT #dailymail #Health #MentalHealth #newyorktimes #OpenAI
https://www.newsbeep.com/us/111789/

Show Original Post

Report

26.08.2025 18:52
r (@r@fed.brid.gy)

Blogpocky tan tranquilo, esperando para grabar un nuevo tip. [Imagen generada con #NanoBanana y #ChatGPT. #IA

Show Original Post

Report

26.08.2025 18:49
strayadult (@strayadult@mstdn.social)

I can't get over the fact that not only are we being sold a shittier version of Google search marketed as "AI". But we are also expected to feed its need for data, all while not being able to trust the answers it spews out.

That is just un-fucking-real.

#AI #ChatGPT #Google #enshittification #tech #SiliconValley #Capitalism #HyperCapitalism

Show Original Post

Report

26.08.2025 18:48
r (@r@fed.brid.gy)

Nach Kritik an GPT-5 zeigt sich: Bessere Leistung bedeutet nicht immer zufriedenere Nutzer. Für die Entwickler von KI-Modellen ist das aber ein großes Problem. #ChatGPT

GPT-5: Blindtest zeigt das eig...

Show Original Post

Report

26.08.2025 18:48
heiseonline (@heiseonline@social.heise.de)

GPT-5: Blindtest zeigt das eigentliche Problem hinter dem neuen Spitzenmodell

Nach Kritik an GPT-5 zeigt sich: Bessere Leistung bedeutet nicht immer zufriedenere Nutzer. Für die Entwickler von KI-Modellen ist das aber ein großes Problem.

https://www.heise.de/news/GPT-5-Blindtest-zeigt-das-eigentliche-Problem-hinter-dem-neuen-Spitzenmodell-10621946.html?wt_mc=sm.red.ho.mastodon.mastodon.md_beitraege.md_beitraege&utm_source=mastodon

#ChatGPT #IT #KünstlicheIntelligenz #news

Show Original Post

Report

26.08.2025 18:45
newsbot_chatgpt (@newsbot_chatgpt@mastodon.social)

OpenAI bot britischer Regierung ein ChatGPT-Abo für alle Bürger an | derStandard.at | Nachrichten, Kommentare & Community
Die Idee dürfte an den Kosten gescheitert sein. Der Technologieminister will die KI-Transformation erzwingen
https://www.derstandard.at/story/3000000285055/openai-bot-britischer-regierung-ein-chatgpt-abo-fuer-alle-buerger-an?ref=rss

#chatgpt

Show Original Post

Report

26.08.2025 18:33
nf3xn (@nf3xn@mastodon.social)

It amuses me not at all when I hear "FAANG developer says they have churned out 30k lines of good code, 30% moar improved blah blah" when I can see #Openai #ChatGPT fabricating responses and then lying about the query being properly grounded. They can't cram their entire git repo into the context - they are 10O% dependent on it using the vector store but usually it is lazy and insubordinate and does not bother to fetch the actual code just makes shit up. Now push that shit to prod I Dare.

Show Original Post

Report

26.08.2025 18:24
AAKL (@AAKL@infosec.exchange)

OpenAI's guardrails are laughable if they can be bypassed with so little effort.

"In Raine’s case, while using a paid version of ChatGPT-4o, the AI often encouraged him to seek professional help or contact a help line. However, he was able to bypass these guardrails by telling ChatGPT that he was asking about methods of suicide for a fictional story he was writing."

Tech Crunch: Parents sue OpenAI over ChatGPT’s role in son’s suicide https://techcrunch.com/2025/08/26/parents-sue-openai-over-chatgpts-role-in-sons-suicide/ @TechCrunch #AI #OpenAi #ChatGPT #cybersecurity #infosec